Toward the end of his extraordinary career, Richard Feynman, the theoretical physicist who solved some of the toughest problems associated with understanding elementary particles, offered some advice on how to do scientific research: “The scientist analyzes something like a detective does. Like a detective trying to find out what happened when he wasn’t there, given clues. We are trying to figure out what nature is like from clues given by experiments. We have the clues and we try to figure it out. It is more analogous to detective work than anything else.” (Mlodinow (2003), p.42)
Macro theorists like to think of themselves as scientists, engaged in proper scientific research. For most problems, while unable to conduct controlled experiments, they are provided plentiful clues from the broad array of evidence that has been, for an extended period, systematically gathered on the behavior of the economy. Unlike scientific research, however, modern practice in mainstream macroeconomics permits analysts to pick and choose which facts to recognize. Chari et al. (2009, p.243) are illustrative, arguing that macro theorists properly “put up with the reality that no model can, or should, fit most aspects of the data.” That loose standard of behavior allows many clues to be suppressed because they contradict consensus theory. That deeply flawed Ptolemaic approach, a crucial subject the GEM Blog returns to periodically, falsifies mainstream macro theorists’ self-description as scientists.
The list of important, widely-known facts that modern mainstream macro scholars, even as they seek policy relevance, simply disregard is frightening. First and foremost of course is involuntary job loss, the rational existence of which requires derivation of meaningful wage rigidity. Macro researchers, constructing general-market-equilibrium models of employment instability that are inherently unable to accommodate MWR, must ignore the most important characteristic of actual fluctuations. As noted last week, Robert Lucas instructed his colleagues in the New Classical revolution that, given that forced job separation cannot be accommodated by neoclassical modeling, it should be simply set aside: “Involuntary unemployment is not a fact or a phenomenon which it is the task of theorists to explain.” (Lucas (1981), p.243) Surely only really poor detectives knowingly ignore the most essential clue.
Other significant, interrelated facts that are conveniently ignored include nominal demand disturbances that produce significant same-direction changes in employment and output; chronic, time-varying wage rents paid by large establishments; the rationing of LEV hours and jobs, pushing many workers off their neoclassical supply schedule; the concentration of temporary layoffs in large establishments; relatively long LEV versus relatively short SEV job tenure for employees in routinized jobs; the interrelated U.S. progression (beginning in the early 1970s) of stagflation to increased inter-industry wage dispersion to the increased incidence of job downsizing and wage givebacks to the disappearance of stagflation; the critical importance of reference standards in job satisfaction; the tit-for-tat response to job dissatisfaction, indicated in the Ultimatum Game and other experiments of behavioral economists; and the existence and nature of LEV human resource departments, almost all of the functions of which reflects substantial firm investment in constructing and maintaining workplace (equity-based) mechanisms of exchange and relatively little of which reflects employee recruitment. Insufficient attention is also paid to powerful evidence that consumption is largely driven by income and wealth, the practitioner consensus that investment is largely driven by pure-profit expectations, and powerful stock-market indications that investor confidence exerts independent influence on total spending, especially during periods of macro uncertainty rooted in unconvincing stabilization-authority credibility.
The modern macro academy’s practice of setting aside evidence that cannot be accommodated in the consensus general-market-equilibrium model class has deep roots in scientific research gone bad. The Ptolemaic name was first attached to the embarrassing Rube-Goldberg models that were constructed in the centuries-long research effort to defend the consensus (albeit obviously wrong) helio-centric model of the solar system. The macro academy may be engaged in the the most egregious application of the Ptolemaic approach in serious research in the world today. During the next several weeks, the GEM Blog will do the necessary dirty work of illustrating that odious practice of putting evidence behind preservation of reputations and human capital. The profession needs our own Richard Feynman.
Blog Type: New Keynesians Paris, France